Nurali Virani

Sensor fusion for fault detection and classification in distributed physical processes

This paper proposes a feature extraction and fusion methodology to perform fault detection & classification in distributed physical processes generating heterogeneous data. The underlying concept is built upon a semantic framework for multi-sensor data interpretation using graphical models of Probabilistic Finite State Automata (PFSA).While the computational complexity is reduced by pruning the fused graphical model using an information-theoretic approach, the algorithms are developed to achieve high reliability via retaining the essential spatiotemporal characteristics of the physical processes. The concept has been validated on a simulation test bed of distributed shipboard auxiliary systems.

Dynamic data driven sensor array fusion for target detection and classification

Target detection and classification using unattended ground sensors (UGS) has been addressed in literature. Various techniques have been proposed for target detection, but target classification is a challenging task to accomplish using the limited processing power on each sensor module. The major hindrance in using these sensors reliably is, that, the sensor observations are significantly affected by external conditions, which are referred to as context. When the context is slowly time-varying (e.g., day-night cycling and seasonal variations) the usage of the same classifier may not be a good way to perform target classification. In this paper, a new framework is proposed as a Dynamic Data Driven Application System (DDDAS) to dynamically extract and use the knowledge of context as feedback in order to adaptively choose the appropriate classifiers and thereby enhance the target classification performance. The features are extracted by symbolic dynamic filtering (SDF) from the time series of sensors in an array and spatio-temporal aggregation of these features represents the context. Then, a context evolution model is constructed as a deterministic finite state automata (DFSA) and, for every context state in this DFSA, an event classifier is trained to classify the targets. The proposed technique of detection and classification has been compared with a traditional method of training classifiers without using any contextual information.

Context-aware Dynamic Data-driven Pattern Classification ∗

This work aims to mathematically formalize the notion of context, with the purpose of allowing contextual decision-making in order to improve performance in dynamic data driven classification systems. We present definitions for both intrinsic context, i.e. factors which directly affect sensor measurements for a given event, as well as extrinsic context, i.e. factors which do not affect the sensor measurements directly, but do affect the interpretation of collected data. Supervised and unsupervised modeling techniques to derive context and context labels from sensor data are formulated. Here, supervised modeling incorporates the a priori known factors affecting the sensing modalities, while unsupervised modeling autonomously discovers the structure of those factors in sensor data. Context-aware event classification algorithms are developed by adapting the classification boundaries, dependent on the current operational context. Improvements in context-aware classification have been quantified and validated in an unattended sensor-fence application for US Border Monitoring. Field data, collected with seismic sensors on different ground types, are analyzed in order to classify two types of walking across the border, namely, normal and stealthy. The classification is shown to be strongly dependent on the context (specifically, soil type: gravel or moist soil).

Dynamic context-aware sensor selection for sequential hypothesis testing

Dynamic sensor selection rules are obtained based on a context-aware measurement model in the framework of sequential hypotheses testing. The notion of context incorporates the operational conditions that directly affect sensor measurements. While a random context leads to a Bayesian decision rule, an unknown but nonrandom context yields minimax game-based rules. In either case, the resulting sensor selection rule trades off decision performance against the cost of sensor activation and the uncertainty of the true context.

Data-driven robot gait modeling via symbolic time series analysis

This paper addresses data-driven mode modeling and Bayesian mode estimation in hidden-mode hybrid systems (HMHS). For experimental validation in a laboratory setting, an HMHS is built upon a six-legged T-hex robot that makes use of a library of gaits (i.e., the modes of walking) to perform different motion maneuvers. To accurately predict the behavior of the robot, it is important to first infer the gait being used by the robot. The walking robots motion behavior can then be modeled as a transition system based on the pattern of switching among these gaits. In this paper, a symbolic time-series-based statistical learning method has been adopted to construct the generative models of the gaits. Efficacy of the proposed algorithm is demonstrated by laboratory experimentation to model and then infer the hidden dynamics of different gaits for the T-hex walking robot.

Learning context-awaremeasurementmodels

This paper presents machine learning-based measurement models with state-augmenting contexts as a paradigm of dynamic data-driven application systems (DDDAS). In order to formulate well-posed statistical inference problems in realistic scenarios, one needs to identify and take into account all environmental factors and ambient conditions, called contexts, which affect sensor measurements. A kernel-based mixture modeling method carries out this task in an unsupervised manner, and results in a machine-defined context set and a probability distribution on it. The resulting measurement model is guaranteed to have contextual awareness, in the sense that the measurements are mutually independent conditioned on the system state and context. Numerical examples illustrate how contextual awareness improves inference performance in the setting of sequential target detection.

Symbolic analysis-based reduced order Markov modeling of time series data

This paper presents a technique for reduced-order Markov modeling for compact representation of time-series data. In this work, symbolic dynamics-based tools have been used to infer an approximate generative Markov model. The time-series data are first symbolized by partitioning the continuous measurement space of the signal and then, the discrete sequential data are modeled using symbolic dynamics. In the proposed approach, the size of temporal memory of the symbol sequence is estimated from spectral properties of the resulting stochastic matrix corresponding to a first-order Markov model of the symbol sequence. Then, hierarchical clustering is used to represent the states of the corresponding full-state Markov model to construct a reduced-order or size Markov model with a non-deterministic algebraic structure. Subsequently, the parameters of the reduced-order Markov model are identified from the original model by making use of a Bayesian inference rule. The final model is selected using information-theoretic criteria. The proposed concept is elucidated and validated on two different data sets as examples. The first example analyzes a set of pressure data from a swirl-stabilized combustor, where controlled protocols are used to induce flame instabilities. Variations in the complexity of the derived Markov model represent how the system operating condition changes from a stable to an unstable combustion regime. In the second example, the data set is taken from NASAs data repository for prognostics of bearings on rotating shafts. We show that, even with a very small state-space, the reduced-order models are able to achieve comparable performance and that the proposed approach provides flexibility in the selection of a final model for representation and learning.

Analysis of Filtered Thermal-fluid Video Data from Downward Facing Boiling Experiments

During severe accidents in a nuclear power plant, in-vessel cooling may be required to mitigate the risk of vessel failure in the event of core meltdown and subsequent corium contamination. This cooling technique, known as in-vessel retention (IVR), entails flooding the reactor cavity with water. If the temperatures are sufficiently high, IVR may cause downward facing boiling (DFB) on the outer surface of the reactor pressure vessel (RPV), which gives rise to two-phase thermal-hydraulic phenomena. The regimes in DFB may range from film boiling to nucleate boiling, where the efficiency of cooling varies immensely between these two. In the DFB geometry under consideration (i.e., a hemispherical vessel), the collected signals/images are heavily contaminated by unavoidable noise and spurious disturbances, which hinder the extraction of pertinent information, such as film thickness and the boiling cycle. This paper proposes a wavelet-based filtering of sensor measurements for denoising of the nonstationary signals with the future objective of estimating the thickness of vapor films in real time, as needed for process monitoring and control. The proposed concept has been validated with experimental data recorded from a pool boiling apparatus for physics-based understanding of the associated phenomena. [DOI: 10.1115/1.4039470]

Robust adaptive motion planning in the presence of dynamic obstacles

Usually in game theoretic formulations for robust motion planning, the model as well as the capabilities (input set) of all dynamic obstacles are assumed to be known. This paper aims to relax the assumption of known input set by proposing a unified framework for motion planning and admissible input set estimation. The proposed approach models every dynamic obstacle as an uncertain-constrained system and then uses the uncertainty estimation technique to estimate the bounds of those uncertainties. The RRT* algorithm with uncertainty estimation for robust adaptive motion planning in presence of dynamic obstacles is presented in this paper. Simulation examples have been used to validate the proposed algorithm.

Spatiotemporal information fusion for fault detection in shipboard auxiliary systems

This paper addresses the issues of data analysis and sensor fusion that are critical for information management leading to (real-time) fault detection and classification in distributed physical processes (e.g., shipboard auxiliary systems). The proposed technique utilizes a semantic framework for multi-sensor data modeling, where the complexity is reduced by pruning the sensor network through an information-theoretic (e.g., mutual information-based) approach. The underlying algorithms are developed to achieve high reliability and computational efficiency while retaining the essential spatiotemporal characteristics of the physical system. The concept is validated on a simulation test bed of shipboard auxiliary systems.